Amazon OpenSearch Service changelog
86 announcements since February 2025 · 3 GA · 1 price change · 5 Region expansions
Free. One email a day, only on days when something you follow has an announcement, or one a week if you switch it in your preferences. You confirm your email with a one-time code; there is no password.
RSS feed for Amazon OpenSearch Service
October 20265
September 20261
August 20266
17 new insights diagnose root causes of Red and Yellow cluster status
Added in AWS GovCloud (US-East) and GovCloud (US-West), bringing total to 19 RegionsRegion
Automatic semantic enrichment extends to VPC-enabled domains in 11 Regions, OpenSearch 2.19 or later
Collection groups support up to 10,000 collections, up from 1,500
Extended Support patch coverage extended to November 7, 2027; support dates set for more versionsPrice
Network access controls using aws:SourceVpce, aws:SourceVpc and aws:SourceIp condition keys
July 20264
Supports OpenSearch 3.7 with 1-bit scalar quantization, search relevance and Query Insights updates
One-click migration of tenants and saved objects from legacy OpenSearch Dashboards
Agent Toolkit for AWS integration adds curated skill for AI coding agents via AWS MCP server
New log analytics engine with up to 4x better price-performance and 70% lower storage
June 20264
May 20264
April 20266
Rollback of service software updates within 15 days using blue/green deployment
Derived Source skips storing source fields to reduce collection storage, enabled per index
Zstandard codec for index compression, up to 32% smaller than LZ4
Graviton4 storage optimized instances up to 18xlarge and 45 TB, in 10 Regions
March 20269
Agentic chat, investigation agent and agent memory for log analytics in OpenSearch UIGA
Accessible in the AWS Console and publishes insights as EventBridge events
Supports OpenSearch 3.5 with agentic memory, context management and relevance workbench additions
Cross-account data access to OpenSearch domains in the same Region, public or VPC
In-place volume increases extended to volumes above 3 TiB without blue/green deployment
Capacity Optimized blue/green option falls back to batch updates when capacity is short
Adds Amazon Managed Service for Prometheus as a sink for metrics pipelines
Single pipeline accepts OpenTelemetry logs, metrics and traces through a unified endpoint